Empirical Study Of Generating Scenarios In Multi-period Financial Asset Allocation

The generating scenarios in multi-period financial asset allocation is studied in this paper. Event tree based sampling is improved and limits method,moments matching method,clustering analysis method and muti-target event tree generating method are put forward. Taking the future's uncertainties into account, the methods we proposed were applied in forecasting HuShen 300 index and Fund An Xin to have empirical studies against the background of domestic current economic situation. Comparing the results we get from forecasting HuShen 300 index with three generating scenarios'methods and comparing the results we get with improved stratified random sampling and Kim's method, we get the research conclusions. The generating scenarios is studied from the following aspects:At first, the previous researches on generating scenarios and asset allocation have been summarized and reviewed.Secondly, we design several methods to generate binomial event tree, such as limits method, and propose a method to sove multi-target generating scenarios problems.Thirdly, we set forth the improved stratified random sampling based on Kim's method.Fourthly, the methods we proposed were applied in forecasting HuShen 300 index and Fund An Xin to have empirical studies.